HackStack

Project Info

News.ai

Devpost

Inspiration

Reading the news helps people expand their knowledge and broaden their horizons. However, it can be time-consuming and troublesome to find quality news articles and read lengthy, boring chunks of text. Our goal is to make news accessible to everyone. We provide concise, digestible news summaries in a conversational manner to make it as easy as possible for anyone to educate themselves by reading the news.

What it does

News.ai provides a concise and digestible summary of a quality article related to the topic you care about. You can easily ask follow-up questions to learn more information from the article or learn about any related concepts mentioned in the article.

How we built it

We used React.js and Flask for our web app. We used NewsAPI to recommend the most updated news based on preferences. We used Monster API's OpenAI-Whisper API for speech-to-text transcription. We used Monster API's SunoAI Bark API for text-to-speech generation. We used OpenAI's GPT 4 API large language model (LLM) to provide summaries of news articles.

Challenges we ran into

We ran into the challenge of connecting multiple parts of the project. Because of its inherent complexity and interconnectivity, making different APIs and frontend plus backend to work together has been our most difficult task.

Accomplishments we're proud of

We're happy that we established a strong pipeline of API calls using AI models. For example, we converted the user's audio input to text using Whisper API before generating text in response to the user's request using GPT API and finally, we converted the generated text to audio output using Bark API. We are also proud to have integrated the NewsAPI in our recommendation system so we can display the latest news for each user tailored to their preferences.

What we learned

Each of our team members had a deep understanding of a specific part of our tech stack; whether that be the frontend, backend, or usage of AI/LLM models and APIs. We learned a lot about how these tools can be integrated and applied to solve real-world problems. Furthermore, by spending the first day going booth to booth and speaking individually to every sponsor, we learned about the intricacies of each platform and API. This allowed us to build a platform that synthesized the strengths of various tools and technologies. For example, we were able to take advantage of the ease and scalability of Monster API's Whisper and Bark APIs.

What's next

Moving forward, we hope to allow for more personalized search of news articles beyond generic topics. Furthermore, we hope to collect additional personalized characteristics that improve the podcast content and understanding for users.

Analysis

Compare with all teams

View

Metric

Figures cover GitHub contributors during the hackathon window. A co-authored commit counts in full for each author, so per-member totals add up to more than the whole-team figures.

Technology

Found in codeClaimed only
  • CSSIn code
  • HTMLIn code
  • PythonIn code
  • ReactIn code
  • TypeScriptIn code
  • FlaskClaimed
  • OpenAIClaimed

5 of 7 appear in the indexed code. 2 claimed on Devpost could not be matched to code, which may simply mean the tool leaves no trace in the repository.

AI coding agents

No AI coding agent signals were found in this repository.

Detected from committed agent config files and commit authorship. Absence of a signal is not proof an agent was unused.

Codebase size

Source size

20 KB

Source files

16

Counts recognized source files only; vendored directories, binaries and lockfiles are excluded, so this is smaller than the repository on disk.

0 stars